Bayesian regression analysis using poly-t densities

Poly-t densities are defined by the property that their kernel is a product, or ratio of products, of Student-t kernels. These multivariate densities arise as Bayesian posterior densities for regression coefficients, under a surprising variety of specifications for the prior density and the data generating process. Although no analytical expression exists for the integrating constant and moments of these densities, these parameters are obtained through numerical integration in a number of dimensions given by the number of Student-t kernels in the numerator, minus one. The paper reviews how poly-t densities arise in regression analysis, and summarizes the results obtained for a number of models.     

Über die Existenz überall trennscharfer Tests

Ohne Zusammenfassung     

Öffentliches und Geheimes

In the second half of 2009, the entire media agenda in Germany was dominated by the swine flu, next to the German Federal Parliament elections. Criticism was voiced in the media against the pharmaceutical industry for lobbying the nationwide swine flu vaccination, which was widely considered nothing but a product of the industry’s lobbying power. In order to avoid the usual interviews with lobbyists, we attempted to apply a new methodological framework approach in analysing lobbying processes and key stakeholders in the public sphere—and outside. Crucial to the diversified approach, which involves quantitative content analysis of media coverage and press releases, are interviews with journalists, lobbyists’ network analysis and internet research on biographies. Although a conspiracy could not be conclusively proven, the results do indicate associations that validate the criticism of the industry’s lobbying activities.     

Das Wiederholungsparadoxon

In this paper we discuss how repetition of a single statement affects its perceived credibility. Using an experimental design, our results support previous psychological studies on the “truth-effect”, which have shown that repeated messages are considered more credible than non-repeated messages. In addition, our study sheds light on the unanswered question why the truth-effect vanishes or reverses if statements are repeated too often. Analysis strongly points to a parallel, indirect negative effect, caused by participants showing reactance as a consequence of repetition.     

Kognitive Dissonanz »Revisited«

Using the theory of cognitive dissonance, this study investigates whether users of an online magazine prefer contents that agrees to their attitudes to political issues. In contrast to earlier studies, dissonance is defined issue-specifically rather than by party identification. Moreover, personal relevance is also included. In a two-step data collection, attitudes and relevance assessments were measured first. Participants were asked in a second session to look at an online magazine. Background software recorded the selection of and the time spent with specific contents as participants looked at the magazine. Results show that users spent more time with attitude-consistent information. This was, however, mostly due to persons who had rated issue relevance high. Users with low or moderate rating for issue relevance spent significantly less time with attitude-consistent content and significantly more time with attitude-inconsistent information, as compared with users with high relevance ratings. The influence of issue relevance is discussed as an explanation of contradictory results on cognitive dissonance and media use.     

Demokratische Internet-Öffentlichkeit? Ein Vergleich der öffentlichen Kommunikation im Internet und in den Printmedien am Beispiel der Humangenomforschung

The importance of the internet has increased significantly in recent years. The easy access of this »new« medium sparked many hopes for a democratization of the public sphere and of public debate: Many authors argued that the internet will make actors and interpretations of topics visible to those parts of the broader public who are usually marginalized in »old« mass media. In an empirical analysis, we compared print media coverage and internet communication to assess these claims. For the case of human genome research — i.e. a bioscientific topic that reseived extensive media attention in Germany —, we analyzed which actors were given public standing and which frames came to the fore. In contrast to the claims mentioned above, our findings indicate that the debate in the internet does not differ to a significant extent from the offline debate in print media. We could not find a more egalitarian, democratic communication in the internet.